evaluation of water quality in an agricultural watershed...
TRANSCRIPT
w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 6 8 5 – 3 6 9 6
Avai lab le a t www.sc iencedi rec t .com
journa l homepage : www.e lsev ie r . com/ loca te /wat res
Evaluation of water quality in an agricultural watershedas affected by almond pest management practices
Xuyang Zhanga, Xingmei Liua,b, Yuzhou Luoa, Minghua Zhanga,c,*aDepartment of Land, Air and Water Resources, University of California, Davis, 1 Shields Avenue, CA 95616, USAbInstitute of Soil, Water and Environmental Sciences, Zhejiang University, Hangzhou 310029, ChinacCalifornia Department of Pesticide Regulation, 1001 I Street, Sacramento, CA 95814, USA
a r t i c l e i n f o
Article history:
Received 22 February 2008
Received in revised form
16 May 2008
Accepted 20 May 2008
Available online 24 June 2008
Keywords:
Water quality
Pest management practices
Organophosphate
Pyrethroid
SWAT
* Corresponding author. Department of LandTel.: þ1 530 752 4953; fax: þ1 530 752 5262.
E-mail addresses: [email protected], m0043-1354/$ – see front matter ª 2008 Elsevidoi:10.1016/j.watres.2008.05.018
a b s t r a c t
In the last decade, the detection of organophosphate (OP) pesticides in the San Joaquin
River watershed has raised concerns about water quality. This study examined the influ-
ences of almond pest management practices (PMPs) on water quality. The Soil and Water
Assessment Tool (SWAT) model was employed to simulate pesticide concentration in
water as affected by different PMPs. California Pesticide Use Reporting (PUR) data were
used to investigate PMP use trends. Stepwise regression analysis was performed to test
the correlation between specific PMP use and pesticide concentrations in surface water
and sediment. Our results showed an increasing use of reduced risk pesticides and pyre-
throids on almonds. SWAT simulation over the period of 1992–2005 showed decreases in
OP concentrations in surface water. High OP and pyrethroid use in dormant sprays was
associated with high pesticide concentrations in water and sediment. Almond pesticide
use was proved to have significant impacts on the pesticide load in the San Joaquin River
watershed. The PMP which combines the use of reduced risk pesticides with no dormant
spray was recommended for almond orchard use. This paper presented a novel method
of studying the environmental impacts of different agricultural PMPs. By combining pesti-
cide use surveys with watershed modeling, we provided a quantitative foundation for the
selection of PMPs to reduce pesticide pollution in surface water.
ª 2008 Elsevier Ltd. All rights reserved.
1. Introduction is from December to February, coincides with the rainy season
California produces 99% of the US almond crop, 58% of which
was located in the San Joaquin Valley in 2006 (CDFA and
USDA, 2006). Major pests of almonds include navel orange
worm (NOW), San Jose scale (SJS), peach twig borer (PTB),
and European red and brown mite. Pest management prac-
tices (PMPs), such as the application of organophosphate
(OP) pesticides and oil mixtures during the dormant season,
are considered effective in controlling these pests (Rice
et al., 1972; UCIPM, 1985). However, the dormant period, which
, Air and Water Resources
[email protected] (Mer Ltd. All rights reserved
in California potentially causing off-site movement of OP pes-
ticides to water bodies (Zhang et al., 2005; Bacey et al., 2005;
Guo, 2003; Dubrovsky et al., 1998).
OPs such as diazinon and chlorpyrifos have been routinely
detected in the surface water bodies of the San Joaquin River
(SJR) watershed during the rainy season (Spurlock, 2002;
Domagalski et al., 1997). Studies have indicated that runoff
from orchards is a source of these OPs (Domagalski et al.,
1997). Although aerial drift contributes to off-site movement,
surface runoff is the main pathway by which OP pesticides
, University of California, Davis, 1 Shields Avenue, CA 95616, USA.
. Zhang)..
w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 6 8 5 – 3 6 9 63686
are transported to the SJR (California Regional Water Quality
Control Board Central Valley Region, 2006). Furthermore, re-
cent findings revealed adverse effects of OPs on many species
dependent on surface water quality for survival, such as the
California red-legged frog, California tiger salamander, Delta
smelt and many birds (Miller, 2006; Ross et al., 1999; Spurlock,
2002; Werner et al., 2002). The frequent detection of OP insec-
ticides and their toxicity to species dependent on good water
quality has resulted in increased regulation of OP use.
As a result of these increased restrictions on OP use,
researchers and regulators have been in search of viable alter-
native pest management practices. The California Department
of Pesticide Regulation (Elliott et al., 2004) and UC Statewide
Integrated Pest Management Program (UC IPM) have proposed
several alternative PMPs to the currently used dormant OP
applications (Zalom et al., 1999), such as: (1) no dormant treat-
ment with in-season sprays as needed (no dormant use); (2)
dormant oil alone; (3) bloomtime sprays for peach twig borer;
(4) reduced risk pesticides (e.g. spinosad) as a dormant spray;
(5) conventional non-OP pesticides (e.g. pyrethroids); and (6)
pheromone mating disruption. In addition, to promote the
use of reduced risk pesticides, the US Environmental Protec-
tion Agency (USEPA) Office of Pesticide Programs proposed
a list of more than 170 reduced risk or OP alternative pesticide
products (USEPA, 2006). Despite these efforts, however, little
work has been done to quantitatively evaluate the influence
of the proposed PMPs and EPA-listed products on surface water
quality, especially at watershed scale.
Additionally, there is a lack of literature addressing recent
changes in almond pest management practices and their
influence on surface water quality. Two studies have investi-
gated the historical changes of PMPs. Epstein et al. (2001)
used the Pesticide Use Reporting (PUR) data during 1992–
1997 to investigate the changes in pest management practice
in almond orchards during the dormant season. They found
that the use of OPs decreased both in the acreage treated
and by the percent of growers applying OP products, while
the use of pyrethroid, BT, oil alone, and no dormant treatment
increased. Another study further confirmed the decreasing
trend of dormant OP use by analyzing the PUR data from
1992 to 2000 (Zhang et al., 2005). While these two studies pro-
vided important documentation on the changes of almond
PMPs, they did not, however, cover the more recent changes
in the years since 2000. In addition, none of the past research
has studied the influence of almond PMPs on the environ-
ment, especially on surface water quality. Therefore, the two
objectives of this paper were to (1) identify the use trends of
traditional and alternative PMPs from 1992 to 2005; and (2)
investigate the environmental impacts of the PMPs on surface
water quality.
2. Materials and methods
2.1. Study area
The SJR watershed is an important agricultural production
area located in the Central Valley of California, responsible
for the drainage of 19,000 km2 of primarily agricultural lands
(Fig. 1). There are about 1480 km2 of almond orchards within
the SJR watershed, on which an annual average of 55,728 kg
of OPs was used during 1992–2005 (CDPR PUR database,
2006). About half of the OPs were used in the dormant season.
Many of the pesticides were sprayed on orchards close to
impaired water bodies, which are listed on the Clean Water
303 d list due to the detections of two OP active ingredients,
diazinon and chlorpyrifos. The considerable amount of pesti-
cide use combined with the orchard proximity to the impaired
water bodies (Fig. 1) points to almond orchards as being
important potential sources of agricultural pesticide runoff.
Since the SJR watershed and surface water bodies support
municipal, industrial and agricultural water use in addition
to providing habitat for fish and wildlife species, it is crucial
to reduce the risk to this watershed of OP runoff from almond
orchards.
2.2. Data sources
Pesticide use information from 1992 to 2005 was obtained
from the PUR database maintained by the California Depart-
ment of Pesticide Regulation (CDPR, 2006). The database re-
cords pesticide use information on every application of
a pesticide in production agriculture and also applications
from some non-agricultural entities. The database includes
information on the amount of pesticide product used, the
amount of active ingredient used, the application date, the
planted area and the treated area. The following measures
were used to assess the use trends of PMP. (1) Total reported
amount of applied active ingredient (AI) (kg of AI); (2) total
area treated from all applications even if the same field was
treated more than once (ha treated); and (3) number of almond
growers reporting use of a particular pesticide.
This study acquired other data for watershed modeling
from various databases maintained by federal and state
agencies. Weather data from 1990 to 2005 were obtained
from the California Irrigation Management Information Sys-
tem (CIMIS, 2005). Soil data were obtained from the State
Soil Survey Geographic database (SSURGO) maintained by
the Natural Resources Conservation Service (NRCS, 2006).
Land use data was obtained from EPA’s Land Use and Land
Cover (LULC) spatial data set (USEPA, 2007). Measured data
of stream flow, sediment load and pesticide concentration
were obtained from the National Water Information System
(NWIS, 2007) maintained by USGS for the monitoring site in
the San Joaquin River near Vernalis, California (USGS site ID
11303500).
2.3. PUR data quality
Although the PUR database is the best database available to
reflect pesticide use in California, CDPR has developed error
checking procedures which identify outliers and errors in
variables including rates of use, grower identification and
site location identification (Wilhoit et al., 2001). An error was
identified if the record was considered to be a duplicate or if
the unit for treated areas was anything other than square
feet or acres. Once identified as an error, the record was
deleted. An outlier was identified if the use rate was greater
than (1) 1.12 kg ha�1 (200 lbs per acre) treated; (2) 50 times
the median kg per ha treated for all uses of that product on
Fig. 1 – San Joaquin River Watershed.
w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 6 8 5 – 3 6 9 6 3687
almonds; and (3) a threshold value determined by a neural
network (Wilhoit et al., 2001). Identified outliers were replaced
with the median use rate (kg of AI per ha treated) of the same
product on almond in the same year. For example, if a record
in 1998 was identified as an outlier by the criteria described
above, the use rate of this record denoted as Rate (old) will
be replaced with the median rate of all the records on the
same product in 1998 denoted as Rate (new). Therefore, the
new value of kilograms of AI denoted as kg of AI (new) was cal-
culated using the following equation
kg of AIðnewÞ ¼ kg of AIðoldÞkg of productðoldÞ � RateðnewÞ � treated ha
(1)
The data cleaning processes identified 2386 records as
errors and 2694 as outliers out of total 625,875 records on
almonds from 1992 to 2005. Therefore, errors and outliers
account for 0.81% of the total records queried, indicating
that the raw data in the database are of good quality.
2.4. PMP scenarios
In addition to traditionally used OP and oil combinations, pes-
ticides used by almond growers in the past 14 years include
pyrethroids, reduced risk pesticides (as identified by the
Crop Protection Handbook, Meister and Sine, 2005), Bacillus
thuringiensis (BT) pesticides, carbamates (Carb), pheromones,
and other pesticides that do not belong to any of the previous
groups (see Table S1 in the Supplemental Information). OPs,
listed in descending order by the total amount of use during
1992–2005, included chlorpyrifos, diazinon, azinphos-methyl,
phosmet, methidathion, naled, parathion, malathion, fenami-
phos, methyl parathion ethoprop, dimethoate, acephate,
phosalone, dicrotophos, and disulfoton. Pyrethrins and
pyrethroids (PYs), in descending order of use, included per-
methrin, esfenvalerate, lambda-cyhalothrin, pyrethrins,
cypermethrin, cyfluthrin, and tau-fluvalinate. Reduced risk
pesticides included tebufenozide, methoxyfenozide, difluben-
zuron, spinosad, pyriproxyfen, fenpyroximate and etoxazole,
with tebufenozide and methoxyfenozide being the two most
commonly used AIs. Oils included various petroleum oils,
mineral oils and soybean oils. BTs included all registered
strains and the encapsulated delta endotxoin of B. thuringien-
sis. Carbamates included carbaryl, methomyl, aldicarb and
formetanate hydrochloride. Pheromones included (E )-5-
decenyl acetate, (E )-5-decenol, Z-8-dodecenyl acetate,
E,E-8,10-dodecadien-1-ol, E-8-dodecenyl acetate and Z-8-
dodecenol.
Six almond PMPs were identified from the PUR database,
labeled PMP1–PMP6 (Table 1). PMP1 is defined as the tradition-
ally used dormant application of OPs and oil, PMP2 is a combi-
nation of OPs, PYs, and oil, PMP3 combines solely PYs and oil,
PMP4 uses only oils, PMP5 only uses reduced risk controls, BT,
pheromone, or controls listed under ‘‘other’’, and, finally,
PMP6 eliminates dormant season applications all together,
postponing pest treatment until after the rainy season. If
reduced risk pesticides were used in combination with OP
and PYs, the PMP was categorized into PMP1 (if with OP),
PMP3 (if with PYs) or PMP2 (if with both OP and PYs).
Each of these PMPs has strengths and weaknesses in terms
of effectiveness, economics, and environmental impact.
While PMP1 and PMP2 are effective in controlling PTB, SJS,
and aphids, their environmental impact is a problem due to
the potential runoff of OPs. PMP3 eliminates OPs, but is less
effective in controlling SJS. PMP4, which eliminates both OPs
and PYs, is less effective in controlling both PTB and SJS.
PMP5 uses reduced risk products which often are more expen-
sive due to higher material costs and increased numbers of
Table 1 – Almond pest management practices
Pest managementstrategies
Code Description
OP PMP1 Traditional use of OPs in mixture
with oil; effective for PTB, aphids
and SJS.
OPþ PY PMP2 Uses of pyrethroid in combination
with OP and Oil
PY PMP3 Pyrethroid in mixture with oil, no
OP application; not as effective as
OPs for scale control; effective in
most areas for PTB
Oil alone PMP4 Only use oil to control European
red mite eggs, brown almond mite
eggs and low to medium
populations of San Jose scale; this
PMS cannot control peach twig
borer or webspinning mites
Lower risk PMP5 Uses of reduced risk, BT,
pheromone or ‘‘other’’ pesticides
as defined in Table S1 in the
Supplemental Information
No dormant use PMP6 Did not use any insecticides during
dormant season, growers rely on
monitoring and uses of insecticides
during in-season
w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 6 8 5 – 3 6 9 63688
applications and monitoring. While BT and spinosad are effec-
tive against PTB and moderate levels of SJS, they are less effec-
tive in controlling aphids. Finally, PMP6 eliminates all pest
control during the dormant season, but requires vigilant
monitoring of pest pressure, and may not always be a feasible
option for the grower.
2.5. SWAT model calibration
The Soil and Water Assessment Tool (SWAT) model was used
to simulate the effects of the six PMPs on surface water qual-
ity. This model was developed to predict the impacts of land
management practices on water, sediment and agricultural
chemical yields at watershed scale (Neitsch et al., 2001) and
has been successfully applied to simulate the hydrology of
agricultural watersheds as well as their pesticide and nutrient
loads (Larose et al., 2007; Hu et al., 2007). The SWAT model
uses a number of submodels to predict runoff, pesticide move-
ment, evapotranspiration, sediment transport and nutrient
movement. Seven databases were constructed as inputs:
soil, weather, land use, fertilizer, pesticide, tillage and urban
land use. The output of the model provided information for
predicting the loads of various pesticides on surface water
for each year.
The hydrology and pesticide components of the SWAT
model were accurately calibrated using monitoring data
from the SJR watershed. A detailed description of the calibra-
tion processes was presented in our previous publication (Luo
et al., in press). Daily simulations of the stream flow, sediment
and pesticide loads were calibrated against monitoring data.
The calibrated stream flow, diazinon and chlorpyrifos loads
are shown in Figs. 2 and 3 as examples. The simulation from
the hydrology components matched the actual measure-
ments supplied by the monitoring data quite closely, with
a Nash–Sutcliffe efficiency of 0.96 (Luo et al., in press, Fig. 2).
Although monitoring data for pesticide loads (converted
from concentration) was very limited, the simulated loads
for diazinon and chlorpyrifos appeared to match well with
measurement data from USGS (Fig. 3). The calibrated model
was then used to predict the loads in surface water of the
major OP and PY pesticides: chlorpyrifos, diazinon, permeth-
rin and esfenvalerate, with pesticide use on almonds from
1992 to 2005. Given the non-point source nature of agricultural
water pollution, the simulation was also performed with the
pesticide use data for all the crops within the watershed, in
order to better understand the environmental impacts of
almond PMPs as compared with that of other commodities.
2.6. Data analysis and statistical tests
Based on the timing of applications reported in the PUR data-
base, the PUR data were separated into two seasons: dormant
and in-season. Dormant season was defined as December–
February, while in-season was from March to November.
Data on kilograms of AI and number of treated hectares
were compiled first by each chemical group (OPs, PYs, reduced
risk, BT, pheromone, other) and then grouped into each of the
six PMPs. To understand the influence of the different PMPs on
water quality, a stepwise regression analysis was then used to
test the correlation between PMPs and the chemical concen-
trations predicted by the SWAT model for each season. Chem-
ical concentrations were selected as the dependent variables,
and kg of AI used by each PMP as independent variables. Two
models were chosen to analyze concentrations in water and
sediment, using the four major pesticides (chlorpyrifos, diaz-
inon, permethrin and esfenvalerate), to determine the models
with the best fit. PMP6 was not included in this analysis
because this PMP by definition does not have kg of AI data.
All the data were processed and analyzed using SAS 9.1.2
and Minitab 14.3.
3. Results
3.1. Pesticide use trends
During the dormant season, oil was the pesticide product with
the highest kilograms of active ingredient (1.1� 106 kg per
year), followed by OPs with an average of 2.8� 104 kg per
year. Kilograms of PYs were just below OPs, increasing from
1.0� 103 kg in 2000 to 2.4� 103 kg in 2005. Kilograms of BT
and carbamates decreased dramatically from mid 1990s to
recent years. Reduce risk products showed an increase from
negligible values in 2000 to 2.1� 103 kg of AI in 2005 (Fig. 4).
Analyzing use trends by the number of hectares treated
rather than by total amount offered a slightly different view,
with pyrethroids and OPs changing ranks. While oil again
ranked the highest (4.0� 104 ha per year), PYs exceeded OPs
since 1997 in terms of treated acreage, nearly doubling in
amount from 2000 to 2005 (from 1.0� 104 to 2.1� 104). Hect-
ares treated with OPs ranked third after PY, with little change
over the time period (from 0.4� 104 to 0.7� 104). Similar to the
analysis of total weights, the hectares treated with reduced
risk pesticides showed a gradual increasing trend, ending
0
200
400
600
800
1000
1200
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Flo
w (m
3 s
-1)
SimulatedObserved
Fig. 2 – Monthly flow calibration of the SWAT model. The monthly average of stream flow rate was aggregated from daily
data.
w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 6 8 5 – 3 6 9 6 3689
with 3.8� 104 ha in 2005, which surpassed that of the OP group
in the last year. Very few hectares were treated with BT, carba-
mates, or pheromones (Fig. 5).
In contrast to the chemical group trends, the trends of the al-
mond PMPs experienced more dramatic fluctuations, as seen
over the time period from 1992 to 2005 (Fig. 6). For the dormant
season in 1992, the percent of growers practicing different
PMPs in descending order was PMP1(69.8%)> PMP6
(17.0%)> PMP4(5.3%)> PMP5(4.9%)> PMP3 (2.2%)> PMP2(0.8%);
while in 2005, the order was changed to PMP6(47.3%)>
PMP3(19.3%)> PMP5 (12.9%)> PMP4 (9.9%)> PMP1(9.3%)>
PMP2(1.3%) (Fig. 6). Fig. 6 shows that the use of PMP1 decreased
dramatically by percent of growers, kg of AI, and treated hect-
ares from 1992 to 2002, but increased afterwards. In its place,
PMP6 has become the most popular pest management prac-
tice, in terms of percentage use among growers. In 2005,
47% of almond growers followed PMP6 compared to 17% in
Dia
-50
0
50
100
150
200
250
Lo
ad
(kg
/year)
Y
1992 1993 1994 1995 1996 1997 199
1992 1993 1994 1995 1996 1997 199
0
20
40
60
80
100
Lo
ad
(kg
/year)
Cho
Fig. 3 – Calibration of the SWAT model for simulation of diaz
1992. Use of PMP3 increased by percent of growers, kg of AI
and treated hectares since 2000 and became the dominant
pest management practice in 2005, second only to PMP6.
Users of PMP5 decreased from 5% in 1992 to 1% in 2001 but
increased to 13% in 2005. The uses of PMP5 as measured by
kg of AI and treated hectares also increased in recent years.
The uses of PMP2 and PMP4 did not change much over the
years (Fig. 6).
For in-season pest management practices, more growers
used PMP3 than PMP1 in recent years (Fig. 6). The percentage
of growers using PMP3 during the in-season increased
steadily, from 3% in 1992 to 35% in 2005. A dramatic increase
in the use of PMP5 was observed in both kg of AI and treated
hectares. In 2005, PMP5 was the most-used PMP (excluding
PMP6) by kg of AI. Uses of PMP2 increased greatly in recent
years and became the most-used PMP by treated hectares in
2005 (Fig. 6).
zinon
EAR
PredictionObservation
PredictionObservation
8 1999 2000 2001 2002 2003 2004 2005
8 1999 2000 2001 2002 2003 2004 2005
rpyrifos
inon and chlorpyrifos. Adapted from Luo et al. (in press).
Fig. 4 – Use of pesticides on almonds by amount of AI during dormant season. OP: organophosphate; PY: pyrethroids; Carb:
Carbamates; Reduced Risk: Reduced risk pesticides.
w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 6 8 5 – 3 6 9 63690
3.2. SWAT simulation outputs
SWAT simulation results clearly showed the temporal trends
of pesticide concentrations in surface water and sediment.
During 1992–2005, water concentrations decreased from
62.2 ppm to 3 ppm for diazinon and from 8.8 ppm to 1.2 ppm
for chlorpyrifos (Fig. 7). Sediment concentrations of these two
chemicals decreased from 15.1� 103 ppm to 0.6� 103 ppm for
diazinon and from 73.7� 103 ppm to 3.0� 103 ppm for chlor-
pyrifos during the same time period (Fig. 7). Both water and
sediment concentrations were lowest in 2003 but increased
in 2002 and 2004 (Fig. 7).
Fig. 7 also shows that permethrin sediment concentrations
peaked at 2.28� 103 ppm in 1998 and 2.56� 103 ppm in 2002,
and then decreased to 0.07� 103 ppm in 2003. Esfenvalerate
05
10152025303540455055
x 103
Treated
A
reas (h
a)
1992 1993 1994 1995 1996 1997 1998
Fig. 5 – Use of pesticides on almonds by treated areas during do
carbamates; Reduced Risk: reduced risk pesticides.
sediment concentrations reached the maximum level of
1.18� 103 ppm in 1998 and decreased to 0.13� 103 ppm in
2003. Water concentrations of these two chemicals showed
a similar pattern having maximum values in 1998 and 2002.
Both water and sediment concentrations increased from
2003 to 2005.
Table 2 shows the percent of use reduction needed to meet
federal and state water quality standards. Current chlorpyrifos
use exceeded EPA and CDFA standards by 2.96% and 11.8%,
respectively. Diazinon use exceeded EPA and CDFA standards
by 1.1% and 2.7%, respectively. To meet the EPA’s standard of
41 ng l�1 and the CDFA’s standard of 15 ng l�1, chlorpyrifos
use needed to be reduced by 88.4% and 95.8% in the SJR water-
shed. Reductions of 61.6% and 78% in diazinon use were
needed to meet the EPA and CDFA standards (Table 2).
Oil OP PYBT Carb Reduced Risk
1999 2000 2001 2002 2003 2004 2005
rmant season. OP: organophosphate; PY: pyrethroids; Carb:
PMP1 PMP2 PMP3PMP4 PMP5 PMP6
020406080
100120140160
0
10
20
30
40
50
60
0
10
20
30
40
50
60
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
Percen
tag
e o
f g
ro
wers
0102030405060708090
Treated
A
reas (h
a)
0
200
400
600
800
1,000
1,200 x103
x103
x103
x103
Am
ou
nt o
f A
I (kg
)
Dormant In-season
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
Fig. 6 – Use of different PMPs in almond by percentage of growers, amount of AI, and treated areas during dormant and
in-season.
w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 6 8 5 – 3 6 9 6 3691
3.3. Impacts of PMP on water quality
The results from the stepwise regression analysis were shown
in Table 3. Only significant correlations were reported. Dor-
mant use of PMP1 was positively correlated with diazinon
and chlorpyrifos concentrations in water and sediment. It
accounted for 72.5% and 78.3% of the variations of diazinon
and chlorpyrifos concentrations in water, and 59.7% and
66.2% in sediment (Table 3). In addition, the use of PMP1 was
negatively correlated with permethrin concentration in water
and sediment contributing to 12.3% and 11.0% of the varia-
tions, respectively. Dormant use of PMP2 was positively asso-
ciated with permethrin and esfenvalerate concentrations in
water and sediment. It accounted for 47.4% and 38.8% of the
variations of permethrin and esfenvalerate concentrations
in water, and 45.9% and 34.7% in sediment. In contrast, dor-
mant uses of PMP3 were negatively correlated with diazinon
concentrations in sediment (5.5%), while PMP4 was negatively
correlated with diazinon in water (4.5%). Finally, PMP5 had
negative correlations with diazinon and chlorpyrifos concen-
tration in both water and sediment, explaining 8.5% and 9.9%
of variations of diazinon and chlorpyrifos in water, and 13.0%
and 6.8% for each chemical in sediment (Table 3).
Most in-season PMPs were not significantly correlated with
any of the independent variables (Table 3). Of those that were
significant, in-season use of PMP1 was negatively correlated
with permethrin in water (13.8%) and sediment (11.5%), while
in-season use of PMP2 had a positive correlation (17.4%) with
esfenvalerate water concentration. All the models were statis-
tically significant with P-values less than 0.05 (Table 3).
0
5
10
15
20
25
Co
ncen
tratio
n in
w
ater (m
g l-1)
X 10-3 Permethrin
Esfenvalerate
0
500
1000
1500
2000
2500
3000
Co
ncen
tratio
n in
sed
im
en
t
(m
g kg
-1)
PermethrinEsfenvalerate
0
10
20
30
40
50
60
70
80 X 103
Co
ncen
tratio
n in
sed
im
en
t
(m
g kg
-1)
DiazinonChlorpyrifos
0
10
20
30
40
50
60
70
Co
ncen
tratio
n in
w
ater
(m
g l-1)
DiazinonChlorpyrifos
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
Fig. 7 – Simulated concentrations of pesticides in sediment and water.
w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 6 8 5 – 3 6 9 63692
3.4. The contribution of almond pesticide use to surfacewater pesticide load
Almond pesticide use had a significant contribution to total
pesticide loads in the watershed (Fig. 8). The pesticide loads
at the watershed outlet predicted by the total pesticide use
of all commodities in the regions were dominated by the loads
predicted from pesticide use on almonds. The temporal trends
of pesticide load predicted by almond pesticide use synchro-
nized with that predicted by total commodity use for both
OP and pyrethroid pesticides. In terms of load, majority of
diazinon and chlorpyrifos mass were associated with water
column while majority of permethrin and esfenvalerate
mass were associated with sediment. Therefore, we ran our
correlation test for the OP and PY pesticides in water and
sediment, respectively. Diazinon loads in water contributed
Table 2 – Use reductions to meet water quality standards
No exceeding (%) 5% e
Chlorpyrifos EPA (41 ng l�1)a 88.4
CDFA (15ng l�1)a 95.8
Diazinon EPA (170 ng l�1)a 61.6
CDFA (100 ng l�1)a 78
a The EPA and CDFA criteria are four day average according to CVRWCB (
tration from daily simulation of SWAT model and the chronic water qua
by almond pesticide use were highly correlated with the total
pesticide loads of all commodities, as can be seen by an R2
value of 0.956, while the correlation for chlorpyrifos was not
high with an R2 value of 0.533 (Fig. 8). Predicted permethrin
and esfenvalerate loads in sediment were also highly corre-
lated with almonds and total commodities, with an R2 value
of 0.93 and 0.95, respectively (Fig. 8).
4. Discussion
The presence of pesticides in surface water poses potential
risks to both aquatic species and human (Arias-Estevez
et al., 2008). As an important source of pesticide input to sur-
face water, agricultural pesticide use strategies should be
carefully examined to evaluate their impacts on water quality.
xceeding 10% exceeding Remark (%)
– – Exceeding of current use: 2.96
46.4 1.1 Exceeding of current use: 11.8
– – Exceeding of current use: 1.1
– – Exceeding of current use: 2.7
2006). Percent of exceeding was calculated based on pesticide concen-
lity standards.
Ta
ble
3–
Perc
en
to
fv
ari
ati
on
inch
em
ica
lco
nce
ntr
ati
on
ex
pla
ined
by
pest
ma
na
gem
en
tp
ract
ices
Ch
em
ica
lco
nce
ntr
ati
on
Med
iaD
orm
an
tp
est
ma
na
gem
en
tp
ract
ices
In-s
ea
son
pest
ma
na
gem
en
tp
ract
ices
sta
tist
ics
Sta
tist
ics
PM
P1
PM
P2
PM
P3
PM
P4
PM
P5
PM
P1
PM
P2
PM
P3
PM
P4
PM
P5
R2
(%)
P
OP
OPþ
PY
PY
Oil
alo
ne
Lo
wer
risk
OP
OPþ
PY
PY
Oil
alo
ne
Lo
wer
risk
Dia
zin
on
Wa
ter
72.5
(þþþ
)4.5
(�)
8.5
(�)
85.5
<0.0
01
Sed
imen
t59.7
(þþþ
)5.5
(�)
13.0
(��
)78.3
0.0
01
Ch
lorp
yri
fos
Wa
ter
78.3
(þþþ
)9.9
(�)
88.2
<0.0
01
Sed
imen
t66.2
(þþþ
)6.8
(�)
73.0
0.0
01
Perm
eth
rin
Wa
ter
12.3
(��
)47.4
(þþ
)13.8
(��
)73.6
0.0
03
Sed
imen
t11.0
(��
)45.9
(þþ
)11.5
(��
)68.4
0.0
07
Esf
en
va
lera
teW
ate
r38.9
(þþ
)17.4
(þþ
)56.2
0.0
11
Sed
imen
t34.7
(þþ
)34.7
0.0
08
Perc
en
to
fv
ari
ati
on
(%)¼
seq
uen
tia
lsu
mo
fsq
ua
res
of
on
ev
ari
ab
le/t
ota
lv
ari
an
ce.
(þ):
Po
siti
ve
corr
ela
tio
n.
(�):
Nega
tiv
eco
rrela
tio
n.
Th
en
um
ber
of
‘‘þ
’’a
nd
‘‘�
’’sh
ow
the
stre
ngth
of
corr
ela
tio
n.
(þþþ
):P
erc
en
to
fv
ari
an
ce>
50%
;(þþ
)o
r(��
):p
erc
en
to
fv
ari
an
ceb
etw
een
10%
an
d50%
;(þ
)o
r(�
):p
erc
en
to
fv
ari
an
ce<
10%
.
w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 6 8 5 – 3 6 9 6 3693
The importance of analyzing pesticide use data and alterna-
tive pest management practices for risk assessment and
effective environmental decision-making was emphasized in
a recent review (Arias-Estevez et al., 2008). This paper pre-
sented the potential of combining the pesticide use surveys
and watershed modeling in the adoption of lower risk PMPs
to replace traditional ones.
The results of this study not only present an informative
snapshot of the environmental impacts on surface water
quality by almond pest management practices but also offer
a basis for recommendations to further reduce impact. The
regression analysis confirmed that PMPs including organo-
phosphates and pyrethroids (PMP1, PMP2) are positively
associated with their concentrations in water and sediment.
While PMP3 did not have any significant association with
pyrethroid concentrations, it is probable that it will in the
future if PMP3 continues to increase in popularity. In con-
trast, the analysis showed that the PMPs using solely oils
(PMP4) or reduced risk products (PMP5) had either negative
association or no association with concentrations of the
four representative chemicals. These results are highly rele-
vant given impending future policy to restrict use of OP and
PY pesticides due to their negative impacts on various non-
target groups.
The analysis of trends on PMP use shows an optimistic
future for reducing impact on water quality. PMP1 has dra-
matically decreased over time in terms of the percentage
of growers employing it, the total kilograms of active in-
gredients, and the total hectares treated, while PMP2
ranked very low by all these measurements. PMP3 is the
greatest concern as its use has been increasing, which
could eventually increase concentrations of pyrethroids
in water systems. However, recent awareness of the
potential adverse effect of pyrethroids on aquatic species
may result in tighter regulations on pyrethroid products,
affecting growers’ preferences for pyrethroid PMPs in the
future. The EPA’s recent decision to re-evaluate over 600
pyrethroids confirmed this surmise. While PMP4 remained
relatively unchanged, PMP5 experienced great increases in
use in recent years. Use of PMP6, which advises no dor-
mant spraying and therefore causes the least negative
environmental impact, has been dramatically increasing
over time. Monitoring pest pressure allowed growers to
take timely action to control pests and to avoid the need
for blanket application of pesticides. The increasing use
of PMP6 indicated that no dormant application was neces-
sary for most of the almond orchards, although during
years with high pest pressure, chemical controls were still
needed. The decreased use of PMP1 together with the sig-
nificant increased use of PMP5 and PMP6 indicates a prom-
ising future for reducing adverse impacts of almond pest
management practices on water quality.
These results are even more promising when taken into
consideration under the larger framework of agricultural
non-point source pollution. The contributions to pesticide
loads from almonds make up a significant proportion of the
total load created by all commodities in the watershed. There-
fore, the positive trends in reducing environmental impact
seen by almond growers will go a long way toward reducing
the overall non-point source pollution in this region. Almond
Chlorpyrifos loadin water
y = 2.1809x + 0.2079R2 = 0.5330
0.0
0.5
1.0
1.5
2.0
2.5
0.0 0.2 0.4 0.6 0.8 1.0
Load by almond pesticide use (mg/s)
Load by almond pesticide use (mg s-1) Load by almond pesticide use (mg s-1)
Load by almond pesticide use (mg s-1)
Diazinon load in water y = 1.3952x + 0.1971R2 = 0.9561
0.001.002.003.004.005.006.007.00
0.00 1.00 2.00 3.00 4.00
Permethrin load in sedimenty = 1.214x + 0.0014
R2 = 0.9299
0.000.010.020.030.040.050.060.070.08
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07
Esfenvalerate load in sedimenty = 2.5804x + 8E-05
R2 = 0.9495
0.0000.0010.0010.0020.0020.0030.0030.0040.0040.0050.005
0.0000 0.0005 0.0010 0.0015 0.0020Lo
ad
b
y all p
esticid
e u
se
(m
g s
-1)
Lo
ad
b
y all p
esticid
e u
se
(m
g s
-1)
Lo
ad
b
y all p
esticid
e u
se
(m
g s
-1)
Lo
ad
b
y all p
esticid
e u
se
(m
g s
-1)
Fig. 8 – Correlation between pesticide loads predicted by uses of the pesticide on almonds and all crops.
w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 6 8 5 – 3 6 9 63694
orchards near waterways should be carefully managed with
low risk pest management practices to minimize pesticide
runoff. In orchards with low PWB and SJS pressure, PMP6
may be the best option for reducing impact on water quality.
With careful monitoring and timely application of reduced
risk pesticides during in-season, it is possible to control pests
without any dormant application. It has been demonstrated
that pest damage levels were not higher in the blocks with
no insecticide applications than those in the conventional
blocks which had dormant OP applications (Elliott et al., 2004).
The relationships between pesticide use and loads in sur-
face water are much more complex than an intuitive thinking
that ‘‘higher use resulted in higher pesticide loads’’. Pesticide
loads not only depend on the amount of use, but also the land-
scape characteristics, soil hydrologic related properties, and
many others. For example, factors such as quantity of pesti-
cide use, physical and chemical properties of pesticides, quan-
tity and patterns of rainfall, and timing of pesticide
application affect whether or not and/or how much pesticides
runoff to surface water bodies. Applications of pesticides in
non-vulnerable areas may not pose environmental risks to
water quality. A recent study conducted in vineyards in
a northwest Spain watershed suggested that soils in these
vineyards were not prone to transport fungicides and;
therefore, fungicide use in these vineyards would not pose
any significant risk to water supplies in the study area
(Bermudez-Couso et al., 2007). Timing of pesticide application
and rainfall also play important roles. Uses of pesticides dur-
ing an irrigation season may not have the same impact as the
uses in a rainy season. Our data showed that OP and pyre-
throid concentrations were lowest in 2003 while the total
uses of these pesticides were not. The relative low concentra-
tions were due to a combination of relatively low dormant
pesticide use and low discharge in year 2003.
The percentage of the total pesticide loads that were
contributed by almonds relative to other commodities in
the area differs among pesticides. The differences could be
explained by variation in application timing of these chem-
icals by almonds in comparison to other crops. Fig. 9 shows
that the application timing of chlorpyrifos on almonds and
other crops were synchronized while the application timings
of permethrin were not. Almonds used the majority of per-
methrin during rainy season while other crops used the
majority of permethrin during dry season (Fig. 9). These
different patterns in application timing may result in differ-
ent contributions of pesticide loads associated with the pes-
ticide use on almonds. Pesticides used in rainy season are
generally more prone to runoff, which explains the large
contribution of almonds to total use of permethrin. There
might be other reasons for the differences in almond contri-
bution patterns among chemicals, such as chemical proper-
ties and transport pathways. Hence, further analysis is
needed.
The use of the SWAT model to predict concentrations
of the four representative pesticides in conjunction with
use of pest management practice data presented a novel
method to study the ecological footprint of growers’ farm-
ing practices on water quality. It appears that almond
growers are well on their way to transitioning to lower
impact systems. However, outreach and education to
growers and their pest control advisors can further the
process through assisting in the implementation of low
impact practices and supplementing chemical choices
with best management practices (BMPs) that can mitigate
and/or prevent negative environmental impact. As
growers and policy makers are better informed as to the
ecological footprints of their actions, they will be able to
better transit to more sustainable practices, and the future
0
10
20
30
40
50
60
70
80
Jan Feb March April May June July Aug Sep Oct Nov Dec
x 103
Mo
nth
ly u
se o
f ch
lo
rp
yrifo
s (lb
s o
f A
I)
Chlorpyrifos use on other cropsChlorpyrifos use on almonds
0
1
2
3
4
5
6
7
Jan Feb March April May June July Aug Sep Oct Nov Dec
x 103
Mo
nth
ly u
se o
f p
erm
eth
rin
(lb
s o
f A
I)
Permethrin use on other cropsPermethrin use on almonds
Fig. 9 – Monthly uses of chlorpyrifos and permethrin by almonds and by all commodities in the watershed.
w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 6 8 5 – 3 6 9 6 3695
of water quality in the San Joaquin River watershed will be
better assured.
5. Conclusion
This paper presented a novel method to study the environ-
mental impacts of different agricultural pest management
practices. By combining the pesticide use surveys and
watershed modeling, we revealed a promising future for
the adoption of lower risk PMPs to replace traditional
ones. While use of traditional dormant OP with oil (PMP1)
decreased dramatically, pyrethroids in combination with
OP and oil (PMP2), pyrethroids in mixture with oil (PMP3),
reduced risk pesticides (PMP5) and no dormant application
(PMP6) have been increasingly used. Although the decrease
of dormant OP use resulted in reduced OP concentrations
in water, the increasing use of pyrethroids in combination
with oil (PMP3) or/and with OP (PMP2) may cause potential
risk to the water quality within the SJR watershed. How-
ever, as reduced risk pesticides and no dormant application
become more and more popular, it is possible to reduce
water quality impact from almond orchards. Since pesti-
cide use on almonds contributes greatly to pesticide loads
in the watershed, especially for permethrin, outreach
efforts are needed to promote better pest management
practices and implement BMPs in almond orchards. This
study suggests an optimistic future of reducing OP pesti-
cide pollution in the SJR by adopting alternative almond
pest management practices.
Acknowledgement
The authors thank the Coalition for Urban/Rural Environ-
mental Stewardship (CURES) and the California State
Water Quality Control Board (CSWQCB) for their financial
support of the research. We also thank Dr. Frank
Zalom for providing materials on his report of ‘‘Alterna-
tives to chlorpyrifos and diazinon dormant sprays’’.
We are very grateful for the reviewer and editor’s
comments.
Supplementary information
Supplementary data associated with this article can be found,
in the online version, at doi:10.1016/j.watres.2008.05.018.
w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 6 8 5 – 3 6 9 63696
r e f e r e n c e s
Arias-Estevez, M., Lopez-Periago, E., Martınez-Carballo, E., Simal-Gandara, J., Mejuto, J.C.c, Garcıa-Rıo, L., 2008. The mobility anddegradation of pesticides in soils and the pollution ofgroundwater resources. Agriculture, Ecosystems andEnvironment 123 (4), 247–260.
Bacey, J., Spurlock, F., Starner, K., Feng, H., Hsu, J., White, J., Tran, D.M., 2005. Residues and toxicity of esfenvalerate and permethrinin water and sediment, in tributaries of the Sacramento and SanJoaquin Rivers, California, USA. Bulletin of EnvironmentalContamination and Toxicology 74 (5), 864–871.
Bermudez-Couso, A., Arias-Estevez, M., Novoa-Munoz, J.C.,Lopez-Periago, E., Soto-Gonzalez, B., Simal-Gandara, J., 2007.Seasonal distributions of fungicides in soils and sediments ofa small river basin partially devoted to vineyards. WaterResearch 41 (19), 4515–4525.
California Regional Water Quality Control Board Central ValleyRegion, 2006. Amendments to the Water Quality Control Planfor the Sacramento River and San Joaquin River Basins for theControl of Diazinon and Chlorpyrifos Runoff into theSacramento–San Joaquin Delta, June 2006. Final Staff Report.Resolution No. R5-2006-0061, Sacramento, CA, USA.
CDFA and USDA (California Department of Food and Agricultureand U.S. Department of Agriculture), 2006. 2006 CaliforniaAlmond Acreage Report. United States Department ofAgriculture National Agricultural Statistics Service,Sacramento, CA, USA.
CDPR (California Department of Pesticide Regulation), 2006.Summary of Pesticide Use Report Data 2005. CaliforniaDepartment of Pesticide Regulation, Sacramento, CA, USA.
CIMIS (California Irrigation Management Information System),2005. Department of Water Resources. Available from: http://wwwcimis.water.ca.gov/cimis/data.jsp (accessed 03.12.05).
Domagalski, J.L., Dubrovsky, N.M., Kratzer, C.R., 1997. Pesticides inthe San Joaquin River, California-inputs from dormant sprayedorchards. Journal of Environmental Quality 26 (2), 454–465.
Dubrovsky, N.M., Kratzer, C.R., Brown, L.R., Gronberg, J.M.,Buro, K.R., 1998. Water quality in the San Joaquin–TulareBasins, California 1992–1995. In: US Geological Survey Circular,vol. 1159. USGS, Sacramento, CA, USA.
Elliott, B., Wilhoit, L., Brattesani, M., Gorder, N., 2004. PestManagement Assessment for Almonds, Reduced-riskAlternatives to Dormant Organophosphate Insecticides.California Department of Pesticide Regulation, Sacramento,CA, USA.
Epstein, L., Bassein, S., Zalom, F.G., Wilhoit, L.R., 2001. Changes inpest management practice in almond orchards during therainy season in California, USA. Agriculture Ecosystems andEnvironment 83 (1–2), 111–120.
Guo, L., 2003. Semi-empirical Prediction of Pesticide Loading inthe Sacramento and San Joaquin Rivers During Winter StormSeason. Environmental Monitoring, EH03-04. CaliforniaDepartment of Pesticide Regulation, Sacramento, CA, USA.
Hu, X., McIsaac, G.F., David, M.B., Louwers, C.A.L., 2007. Modelingriverine nitrate export from an east-central Illinois watershedusing SWAT. Journal of Environmental Quality 36 (4), 996–1005.
Larose, M., Heathman, G.C., Norton, L.D., Engel, B., 2007.Hydrologic and Atrazine simulation of the Cedar CreekWatershed using the SWAT model. Journal of EnvironmentalQuality 36 (2), 521–531.
Luo, Y., Zhang, X., Liu, X., Ficklin, D., Zhang, M. Dynamicmodeling of organophosphate pesticide load in surface waterin the San Joaquin Valley. Environmental Pollution, in press,doi:10.1016/j.envpol.2008.04.005.
Meister, R., Sine, C., 2005. Crop Protection Handbook. MeisterPublishing Company, Willoughby, Ohio, USA.
Miller, J., 2006. Poisoning Our Imperiled Wildlife, San FranciscoBay Area Endangered Species at Risk from Pesticides. Centerfor Biological Diversity, San Francisco, CA, USA. Report.
NWIS (National Water Information System), USEPA, 2007. Availablefrom: http://waterdata.usgs.gov/nwis (accessed 20.12.07.).
Neitsch, S.L., Arnold, J.G., Kiniry, J.R., Williams, J.R., 2001. Soil andWater Assessment Tool-Version 2000 – User’s Manual.Blackland Research and Extension Center, Temple, Texas, USA.
NRCS (National Resources Conservation Service), USDA 2006.Available from: http://www.ncgc.nrcs.usda.gov/products/datasets/ssurgo/ (accessed 20.05.06.).
Rice, R.E., Jones, R.A., Black, J.H., 1972. Dormant sprays withexperimental insecticides for control of peach twig borer.California Agriculture 26 (1), 14.
Ross, L., Hsu, S.J., White, J., Hefner, K., 1999. Distribution and MassLoading of Insecticides in the San Joaquin River CaliforniaSpring 1991 and 1992. The California Department of PesticideRegulation, Sacramento, CA, USA.
Spurlock, F., 2002. Analysis of diazinon and chlorpyrifossurface water monitoring and acute toxicity bioassay data1991–2001. In: P.M.B (Ed.), Environmental HazardsAssessment. California Department of Pesticide Regulation,Sacramento, CA, USA.
UCIPM (University of California Integrated Pest ManagementProgram), 1985. Integrated Pest Management for Almonds,first ed. Division of Agriculture and Natural Resources,University of California, Oakland, CA, USA. Publication 3308.
USEPA (U.S. Environmental Protection Agency), 2006. ReducedRisk/Organophosphate Alternative Decisions for ConventionalPesticides. US Enviromental Protection Agency Office ofPesticide Programs, Washington D.C. USA. Publication 20460.
USEPA (U.S. Environmental Protection Agency), 2007. Availablefrom: http://www.epa.gov (accessed 12.08.07.).
Werner, I., Deanovic, L.A., Hinton, D.E., 2002. Toxicity ofstormwater runoff after dormant spray application ofdiazinon and esfenvalerate (Asana) in a French prune orchard,Glenn County, California, USA. Bulletin of EnvironmentalContamination and Toxicology 68 (1), 29–36.
Wilhoit, L., Zhang, M., Ross, L., 2001. Data Quality of California’sPesticide Use Report. California Department of PesticideRegulation, Sacramento, CA, USA.
Zalom, F.G., Oliver, M.N., Hinton, D.E., 1999. Alternatives toChlorpyrifos and Diazinon Dormant Sprays. Statewide IPMProject. Water Resoruces Center and Ecotoxicology Program,University of California, Davis, CA, USA.
Zhang, M.H., Wilhoit, L., Geiger, C., 2005. Assessing dormantseason organophosphate use in California almonds.Agriculture, Ecosystems and Environment 105 (1–2), 41–58.